AEO Insights
Sourceable·June 10, 2026·6 min read
New AEO data on brand mentions, share of voice, and sentiment across the four major AI answer engines — and what it means for your brand.
ChatGPT dominates AI-referred traffic at 78.23%, but it's only one of four engines now shaping brand discovery (per SE Ranking's 2026 research).
Claude grew 386% in referral traffic in early 2026 — the fastest-growing AI source — and is increasingly an embedded work tool, not just a search box.
Across 240 brands and 1,800 category prompts, the average brand appeared in just 34% of relevant AI answers — meaning most brands are absent from two-thirds of conversations about their own category.
Visibility varied dramatically by engine: a brand named first by Perplexity appeared in only 42% of the equivalent ChatGPT answers.
71% of brand mentions occurred in answers that produced zero referral traffic — influence your analytics can't see.
For twenty years, brand discovery was a search-engine question: where do you rank on Google? That question is no longer enough.
Roughly 58% of searches now end without a click, as AI Overviews and assistants answer on the spot. And a growing share of high-intent research has moved into conversational AI — ChatGPT, Google's Gemini, Anthropic's Claude, and Perplexity. In that world, the unit of discovery isn't a ranked list of ten links. It's a single, confident recommendation naming two or three brands.
This is the core of Answer Engine Optimization (AEO) — sometimes called Generative Engine Optimization (GEO): making sure your brand is the one the AI names. And like early SEO, the brands measuring it now have an enormous head start.
So we set out to measure it.
Using Sourceable, we tracked how 240 brands appear across the four major AI answer engines — ChatGPT, Gemini, Claude, and Perplexity — for 1,800 real, category-defining buyer prompts. For every AI answer, we recorded four things:
Mention — did the brand appear, and how prominently?
Position — how early was it named relative to competitors?
Share of voice — what % of the answer's brand mentions belonged to it vs. rivals?
Sentiment — positive, neutral, or negative framing?
(Full methodology at the end.)
ChatGPT accounts for 78.23% of all AI-referred website traffic, with Claude, Perplexity, Gemini, and Copilot splitting the rest (SE Ranking, 2026).
But here's the catch our data surfaced: traffic share and recommendation share are not the same thing. A brand can be heavily recommended inside an answer — shaping the decision — without ever generating a measurable click. In our set, 71% of brand mentions occurred in answers that produced no referral traffic at all. The influence is real; the analytics are blind to it.
The most striking pattern: AI visibility is not consistent across engines.
A brand that ranked in the top recommendations on Perplexity appeared in only 42% of the equivalent ChatGPT answers. Claude — increasingly used as a workflow tool rather than a search destination — surfaced a different set of brands again, often favoring those with strong structured data and documentation.
Engine | Avg. brand mention rate | Notes |
|---|---|---|
ChatGPT | 41% | Largest reach; broad training recall |
Gemini | 36% | Tied to Google's index + freshness |
Perplexity | 33% | Citation-heavy; rewards linkable sources |
Claude | 29% | Favors structured, well-documented brands |
The lesson: measuring one engine understates your real exposure. If you only check ChatGPT, you're missing where competitors may be quietly winning.
Being mentioned isn't the whole story. In 58% of cases, brands were named with neutral or mixed sentiment — and in 9%, with outdated or inaccurate information the model had absorbed from old sources.
We also tracked citations — whether the AI pointed to the brand's own domain as a source. Brands cited frequently (especially on Perplexity) held their position 2.3× more stably over time, suggesting that owned, machine-readable content is a durable visibility asset.
When an AI recommends a competitor instead of you, nothing in your dashboards moves. No lost referral — the buyer never reached your site. No bounce. No ranking drop — you may still be #1 on Google. The deal simply goes elsewhere, upstream of every report you own.
This is why AI visibility needs its own instrument — the same way Google rankings got rank trackers two decades ago.
Five practical moves, based on what the data shows:
Measure across all four engines, not one. Visibility is inconsistent; a single-engine check is misleading. Sourceable's AI visibility tracker runs that check across ChatGPT, Gemini, Claude, and Perplexity on a schedule.
Benchmark against competitors. Share of voice matters more than raw mentions — being named more and earlier than rivals is the goal.
Strengthen machine-readable signals. Citations correlated with stable visibility; clean structured data, llms.txt, and accurate facts help models represent you correctly.
Watch sentiment, not just presence. An inaccurate or negative mention can hurt more than silence — and it's fixable once you can see it.
Track it over time. AI answers shift. A one-time audit is a snapshot; a weekly trend is a strategy.
We analyzed 240 brands across 1,800 category prompts from March to May 2026, querying ChatGPT, Gemini, Claude, and Perplexity on a recurring schedule via Sourceable. For each response we scored mention, position, competitor share of voice, sentiment, and citations. Macro traffic figures are cited from SE Ranking's 2026 AI-traffic research. Limitation: AI answers are probabilistic; we sampled across multiple runs to produce stable trends rather than single-response snapshots.
Search didn't disappear — it changed shape. The list became an answer, and the answer names a few brands, not ten. Our data is clear: AI visibility is now a distinct, measurable channel — inconsistent across engines, invisible to traditional analytics, and wide open for the brands willing to measure it first.
The brands that treat AI visibility as a metric today — the way they treat Google rankings — will be the default recommendation in their category tomorrow. The rest will keep celebrating rankings while wondering where the pipeline went.
Want to see where your brand stands? Run a free AI visibility check with Sourceable.
Research and analysis by the Sourceable team. Sourceable tracks how AI assistants recommend your brand across ChatGPT, Gemini, Claude, and Perplexity.
Continue reading our latest insights
Ranking #1 on Google means nothing if ChatGPT never names you. The skills that won search are quietly becoming the wrong skills — here's what actually changes when the result is a single generated answer.
AI referral traffic nearly doubled in a year and Claude alone grew 386% in four months. The brands showing up in those answers aren't louder — they're optimized differently. Here's the step-by-step playbook.